CS Wednesday Talks take place in the ASH lobby or auditorium. A light lunch is served at noon. Talks begin at noon and end by 1 p.m.
December 10, 2014
Using Recognition Memory to Encode Information
Brent Heeringa, Associate Professor and Chair of Computer Science at Williams College
Abstract: Whether it is our ability to detect a wide range of colors or to remember a face we haven’t seen for years, the human brain is impressively adept at interpreting, encoding, and remembering visual data. It is not, however, especially adept at remembering strings of arbitrary numbers: the very patterns that identify so much of contemporary life, from credit cards to bank account numbers. This talk discusses an alternative approach that transforms symbolic information into visual representations that are universal and error-correcting.
Biography: Brent Heeringa is associate professor and chair of computer science at Williams College. His research focuses on approximation algorithms and data structures. He enjoys gardening, squash, and indie rock.
December 3, 2014
Family Engagement in Literacy: Translating Research into Practice in Holyoke
Melissa Burch, associate professor of cognitive development
Abstract: Children's experiences in home and school are both valuable contexts to support children's literacy development, and they are most effective when they work together. In this presentation, Melissa Burch and some of the students from her first-year tutorial will discuss the work they have done in partnership with an elementary school in Holyoke. In this talk, we will discuss the context in the school district that surrounds literacy instruction and family engagement. We will also present some of the ways we have supported connections between school and home through Ready to Read (Listo para Leer), a series of newsletters that include family-centered activities to reinforce the five components of reading. Each of the activities was designed to translate research findings into specific practices that children and families could enjoy at home. This project will continue throughout the school year and support other efforts in the district to promote home-school alignment and to support student success.
Biography: Melissa Burch, associate professor of cognitive development, received her B.A. in psychology from Franklin and Marshall College. She earned her Ph.D. in child development with a minor in interpersonal relationships from the University of Minnesota. Melissa's research interests center on memory and narrative development. She has been exploring how parental verbal support may contribute to children’s ability to recall the past as well as how parents can engage children in literacy-related activities to reinforce learning in school.
November 19, 2014
The Technical Sciences and the Purposes of God: Theory and Practice in the Hizmet Movement in Turkey
Caroline Tee, Ph.D., University of Bristol, U.K.
Abstract: This presentation explores the philosophical justification for engagement as religious actors in the technical sciences, showing how practitioners within the movement derive spiritual meaning from the practical application of science, namely in the fields of medicine and engineering, by drawing on the Nursian doctrine of ‘positive action’. This observation is situated within a wider ethnographic framework which traces the activities and evolving priorities of the Hizmet Movement, focusing on its emergence as an actor in the lucrative field of private higher education in Turkey in recent years.
Biography: Dr. Caroline Tee is a postdoctoral research assistant in the department of archaeology and anthropology at the University of Bristol, U.K. She holds an M.A. in Islamic Studies and a Ph.D. in social anthropology. She is currently working on a two-year project funded by The John Templeton Foundation exploring the teaching of science within an Islamic milieu in Hizmet schools in Turkey.
November 12, 2014
Decoding How Neural Systems Process Information About Faces
Ethan Myers, assistant professor of statistics
Abstract: Faces are a biologically important class of stimuli for primates. Recent work has identified six discrete face areas in the temporal lobe of the macaque that form a network which appears to be responsible for processing information related to faces. The vast majority of neurons in these face areas have much higher firing rates to images of faces compared to other object categories, however it is still unclear what types of information, and consequently which visual behaviors, each face area could support. In this work we use neural population decoding analyses to better characterize what information is being represented in three of these face areas (middle lateral/fundus (ML/MF), anterior lateral (AL), and anterior medial (AM) patches). Our decoding results show that there is more information about faces compared to non-face objects in all three regions, and that AM in particular shows a much stronger representation of faces compared to non-face stimuli. Additionally, we find that information about face identity that is invariant to the pose of the head is largely absent from ML/MF, is stronger in AL and is strongest in AM. These findings show that the face patch system builds up visual features in the more anterior patches that are useful for identifying individuals while losing information that is useful for discriminating between different non-face objects.
Biography: Ethan Meyers, assistant professor of statistics, received his B.A. in computer science from Oberlin College, and earned his Ph.D. in computational neuroscience from the Massachusetts Institute of Technology. Ethan’s research focuses on creating methods for analyzing high dimensional neural signals in order to understand the neural processing that underlies object recognition, working memory and other cognitive processes. Ethan’s teaching interest span a range of topics including statistics, machine learning, data science and computational neuroscience.
October 29, 2014
GOAT TALK: Vocal Learning in Mammals
Mark Feinstein, professor of linguistics
Abstract: Productive vocal learning is the ability of an animal to modify its inherited acoustic signals or to develop new sound signals by means of mimicry and other learning processes. It's very familiar to us humans; vocal learning is of course a critical aspect of human language acquisition. A variety of birds also exhibit robust vocal learning ability. But it is quite rare in the biological world at large and is reported for only a handful of non-human mammals including bats, whales and dolphins, seals and sea lions, and elephants. Curiously, there seems to be little if any evidence for productive vocal learning in any of the non-human primates, even the great apes (e.g. chimpanzees) that are so closely related to us. Nor is it found in domesticated mammals like the dog, which live in such close interaction with humans. Dogs can learn to passively recognize and respond to human vocal signals; but they don't learn to systematically modify their own vocalizations to actively "talk" with us or with each other. To the surprise of many, however, there is a humble domesticated ungulate (the goat) that has recently been claimed to be a productive vocal learner. I'll review the vocal learning story, report on the goat study and on related research that I'm undertaking myself, and discuss its evolutionary, behavioral and cognitive implications.
Biography: Mark has taught in the School of Cognitive Science since 1976. Originally trained in linguistics, he did his early research and published work in the study of sound systems in human language (phonology). He soon turned his attention to the biological evolution of speech and language, and his research and teaching focus is now on the comparative study of vocalization, cognition, and behavior in non-human animals. Much of his research has centered on dogs and other canids including the coyote and the New Guinea Singing Dog. Currently he is working on the vocal and social behavior of domesticated animals (and completing a book with Ray Coppinger on the subject); much of his field research has been conducted on free-ranging sheep in the hills of Connemara in the west of Irelan
October 22, 2014
Lightinging Talks by Coppinger Grant Students
Dry in the Sky?: Plant Trait Trade-Offs in Tropical Montane Cloud Forest Canopies across Precipitation Gradients
Erica Hample, Division II Project, with Dr. Sybil Gotsch, in Monteverde, Costa Rica
Abstract: By means of technical tree climbing, we studied the water capacity, foliar uptake, and drought resistance of epiphytes as a model for the health of the Tropical Montane Cloud Forest. We hope to add to the knowledge in predicting responses to projected changes in climate.
The Ghost of Aggression": Patterns of interspecific territoriality in congeneric neotropical songbirds over a major rainfall gradient
Jacob Drucker, Division III Project, Magdalena Valley, Colombia
Abstract: Interspecific territoriality has been shown to be a key factor in determining spatial limits in the distributions of closely related (congeneric) bird species. This behavior is particularly prevalent over habitat gradients, which provide a gradual transition of habitats and consequently bird communities. Here I assess the role interspecific territoriality plays in maintaining three partially overlapping distributions of Thamnophilus antshrikes (Passeriformes, Thamnophilidae) throughout the Magdalena Valley, a major rainfall gradient in Colombia. I present the first evidence of territorial aggression by a smaller congener towards a larger one, as well as aggressive responses to congeners in populations distal from the replacement zone. These findings show that large body size is not always correlated with social dominance, represents conspecific variation in interspecific behavior, and reflects on how behaviors evolved in the Pleistocene and earlier may still show in current generations.
Serotonergic (5-HT) stimulation of enteric neurogenesis: identification of target neuronal phenotypes and 5-HT4 mediation
Garrette Furo, Internship with Dr. Michael Gershon at Columbia University College for Physicians and Surgeons
Abstract: The enteric nervous system (ENS) develops from neural crest-derived precursor cells (ENCDC). These ENCDC give rise to neurons that are born in a phenotype-related sequence. Serotonergic neurons are among the earliest born and coexist with still-dividing ENCDC. As a result, 5-HT from enteric serotonergic neurons can influence the fate of that follow serotonergic neurons in their withdrawal from the cell cycle. We worked on relevant experiments during the summer 2014.
October 15, 2014
Self-Replicating Distributed Virtual Machines
Lance R. Williams, associate professor of computer science, University of New Mexico
Abstract: Recent work showed how an expression in a functional programming language can be compiled into a massively redundant asynchronous spatial computation called a distributed virtual machine. A DVM is comprised of bytecodes reified as actors undergoing diffusion and communicating via messages containing encapsulated virtual machine states. Significantly, it was shown that both the efficiency and the robustness of expression evaluation by DVM increase with redundancy. In the present work, spatial computations that become more efficient and robust over time are described. They accomplish this by self-replication, which increases the redundancy of the elements of which they are comprised. The first and simplest of these self-replicating DVMs copies itself by reflection; it reads itself from a contiguous range of memory. The remainder are quines. As such, they reproduce by translating and transcribing self-descriptions. The nature of the self-descriptions and of the translation and transcription processes differ in each case. The most complex self-replicating DVM described represents a fundamentally new kind of artificial organism--a machine language program reified as a spatial computation that reproduces by compiling its own source-code.
Biography: Lance R. Williams received his B.S. in computer science from the Pennsylvania State University in 1985 and his M.S. and Ph.D. in computer science from the University of Massachusetts at Amherst in 1988 and 1994. His dissertation, in the area of computer vision, was on perceptual completion of occluded surfaces. After completing his Ph.D., he spent four years at NEC Research Institute in Princeton, NJ, where he developed a series of increasingly more general neural models of the process used by the human visual system to compute the shape of object boundaries where they cannot be directly observed. In 1997, Dr. Williams joined the faculty of the department of computer science at the University of New Mexico where he is currently an associate professor. His research since joining UNM has addressed a range of topics in computer vision, neural computation, digital image processing, and artificial life.
October 8, 2014
Division II Concentrations in Cognitive Science: A Presentation by Faculty in CS
Abstract: Are you thinking about Division II? Are you interested in animal behavior, animation, astronomy, cognitive science, computer science, education, game development, journalism, linguistics, media, neuroscience, philosophy, psychology, science and society, or statistics? Would you like to know more about the faculty in CS, their research and teaching interests, and the kinds of Division II projects they have been involved with? Do you like free pizza? Then come to the CS Wednesday Lunch talk on October 8. It will feature very short lightning talks by each of the faculty in CS with lots of time for questions and discussion. And pizza.
October 1, 2014
Towards the Evolution of Adaptive Complexity in Virtual Worlds
Lee Spector, professor of computer science
Abstract: One could presumably re-create the adaptive complexity of the biosphere by finding or creating a planet with the right geology and atmospheric composition, putting it in an orbit at an appropriate distance from a sun-like star, and waiting for about five billion years. A bit of luck might help too. But is there a cheaper, faster way? If we make simplifications and cut corners then might we be able to get "life-like" adaptive complexity to emerge in computer simulations within hours or days? What are the essential ingredients and processes that must be included in such simulations? What can we learn about life and evolution from experiments aimed at creating artificial life? In this talk I will review some of my prior work on these questions and present some of the new directions in which my work in this area is moving.
Biography: Lee Spector holds a Ph.D. in computer science from the University of Maryland and a B.A. in philosophy from Oberlin College. His main interests are artificial intelligence and the connections between cognition, computation, and evolution. He is also interested in the use of technology in music and other arts. His recent research includes projects on the development of new genetic programming techniques, the use of artificial intelligence technologies in the study of quantum computation, the interdisciplinary study of human and machine cognition, and the development of technologies to support inquiry-based education.
September 24, 2014
Wide Minds: How Far Can We Extend Cognition Before It Breaks?
Laura Sizer, dean of the School of Cognitive Science and associate professor of philosophy
Abstract: A family of theories in philosophy and cognitive science argue variously that the mind should be thought of as essentially "embodied," or somehow "extended" out into the world. In other words, the mind is most definitely not all in your head. What does this imply about the nature of cognition? Is my body part of my mind? Is my smart phone? Where does cognition end, and if it doesn't, what does this suggest about the subject matter of cognitive science? This talk explores some of these claims and their implications for cognitive science and philosophy of mind.
Biography: Laura Sizer is the dean of the School of Cognitive Science and associate professor of philosophy. She earned a Ph.D. in philosophy from the University of Wisconsin-Madison. She teaches a variety of courses in the philosophy of mind and cognitive science. Her research currently focuses on affect (emotions and moods), and their relationships to consciousness and cognition.
September 17, 2014
Barking Up the Wrong Tree: the Function of Barking in Dogs
Kathryn Lord, visiting assistant professor of animal behavior
Abstract: The dog bark is one of the most common sounds in our modern world and yet we do not know the function of this vocalization. The dog's bark is unusual in that it is highly variable and expressed in seemingly unending contexts. In this talk I will discuss the two main hypotheses for the function of the dog bark and present the findings of an experiment I conducted with my students last year at Gettysburg College.
Biography: Kathryn Lord holds a Ph.D. in organismic and evolutionary biology from the University of Massachusetts Amherst and received her undergraduate degree from Hampshire College. Her main interests are in the evolution and development of animal behavior, as well as its application to the management of domestic and wild species. This semester she is teaching a course on learning and development in wolves and dogs.
September 10, 2014
Changing Mindsets to Improve Innovation (and Entrepreneurship)
Gillian Marcelle, associate professor of strategy and innovation, University of the Witwatersrand
Abstract: An “intermediary” is a concept in innovation studies to help understand the role of firms, agencies, and individuals that facilitate innovation by providing the bridging, brokering, knowledge exchange necessary to bring together the range of different organizations and knowledge needed to create successful innovation. We lack, however, a deeper understanding of exactly how intermediation impacts innovation. Such an understanding is necessary in order to undertake further research and for the practical purpose of training intermediaries for success. Given the importance of intermediation in ecosystems, how does intermediation impact the innovator and the process of venture creation? We envisage that intermediaries (of various types) engage with innovators (to change their mindset and equip them) and facilitate the development and diffusion of solutions (by providing funding and increasing their viability). We define intermediaries as knowledge and resource brokers (incubators, accelerators, investors, banks, business development support agencies, research councils, universities, and technical colleges) that influence the process of innovation, including the technical design and development of solutions and assisting with moving innovators through various stages of the discovery, venture creation and management process.
Biography: Gillian Marcelle focuses her research and teaching on strategy, innovation, and capability-building. She is an active policy and academic research scholar with more than twenty years’ experience gained in developed and developing country settings. Her research interests include firm-level capability building and learning, industry dynamics: telecoms and IT sector, contemporary South African business and investment climate, corporate social responsibility and social entrepreneurship. She provides academic leadership for the NEPAD Consortium on Science, Technology and Innovation Policy program and supervises masters and doctoral students interested in strategic management of innovation and industry analysis. She teaches courses on strategic management and competitor and industry analysis.